At some point this morning, in London, Federal Reserve Governor Christopher Waller did the thing central bankers almost never do before a print: he named a number. If core consumer prices rise 0.3 per cent or more from September, he said, a 50 basis point hike should be considered at the December meeting. The consensus for that field is 0.2 per cent. The release is at 13:30 UTC, under two hours from this writing. And the CME FedWatch tool, as two readers quote it, assigns 84.4 per cent to a quarter point, 15.6 per cent to no change, nothing to a cut — and nothing at all to the half point the governor just put on the table. The branch he opened is the one the distribution has no mass on.
The trigger sits one tenth above consensus, and one tenth is the entire distance
Three independent readers carry the threshold identically: a Reuters report filed from London and timestamped 09:12 UTC, a second house at 09:09 UTC, and a Tokyo-authored market outlook stamped 10:52. All three give the same conditional — core CPI at 0.3 per cent month on month or above, and a half point becomes live. One further reader, a machine-rewritten flash, carries only the qualitative version, “above forecasts”, with no number; we are not counting it, and we note it because a reader that drops the number while keeping the sentence is exactly how a precise conditional becomes a vague one in three hops.
So the distance from what is expected to what fires the bigger move is 0.1 percentage point, which at the Bureau of Labor Statistics’ published precision for this field is one decimal place. There is no grey zone. Core prints 0.2 and the quarter point stands; core prints 0.3 and a governor has already said in public what he thinks follows.
Yesterday morning this desk published that the stored headline consensus of 0.3 is a rounding of about 0.29, that six named houses ran 0.20 to 0.30, and that the rounding is wider than the disagreement it conceals. Hold that next to this. The named-house spread on the headline field is 0.10 wide. The gap between the core consensus and Waller’s trigger is also 0.10. The branch the market prices at zero sits entirely inside the width of the vendors’ own published disagreement about the adjacent field.
The market has priced the level and not the size
The Federal Open Market Committee held at 3.75 to 4.00 per cent on Wednesday 28 October by ten votes to two, with Cook and Hammack dissenting in favour of a hike. A December move of a quarter point takes the range to 4.00 to 4.25, and that is the outcome the 84.4 per cent is attached to — a specific range, not a direction. Read the two figures together and the market is not saying it is 84 per cent sure the Fed moves. It is saying it is 84 per cent sure the Fed moves by exactly this much, and effectively certain it will not move by twice that.
That is a distributional claim, and it is the one Waller has just made contestable. For a system, the practical consequence is not about direction at all. A position sized against a binary — hike or hold — is sized against the wrong object today, because both of the live outcomes are hikes and they differ by a factor of two in magnitude. The uncertainty that matters has moved from whether to how much, and a December-dated instrument that prices the second at zero will reprice discontinuously if core comes in at 0.3.
We will say the uncomfortable part plainly: this is not a reason to be long dollars into 13:30. It is a reason to notice that the payoff is no longer symmetric around the consensus, and that the asymmetry was put there in public, by a voting member, more than four hours before the number.
What the pair actually did with it: thirty pips
Dollar-yen fell to 157.87 early in London and rebounded to 158.17 after the remarks — a 30.0 pip round trip on the day’s range so far, with the dollar index moving from 101.408 to 101.688, a rise of 0.280 points or 0.276 per cent. Euro-dollar went the other way, from 1.1251 to about 1.1220. Those four figures come from a single market outlook published at 10:52 and are therefore one house’s reading; we have not second-sourced the dollar index levels and say so.
Put it against the session. Tokyo traded 157.89 to 158.12 on the same source, 23.0 pips, against the 157.90 to 158.12 a Tokyo publisher gave to 13:45 Japan time — a one pip difference on a low in a window that was still open when the earlier reading was taken, which is an extension rather than a conflict. The day’s range as that page recorded it at 10:52 is 157.87 to 158.17, 30.0 pips — a window that was still open when it was written, so read it as a floor and not a total. A sitting governor publicly doubled the tail of the December distribution and the pair paid thirty pips for it.
Above, the 200-day remains the ceiling nobody has traded through. Today adds an eighth reading: one house puts it near 158.69, and a bank quoted on another page calls the 158.70 area the level the market is reluctant to push through, coinciding with the 200-day. Our band is now 158.60 to 158.70, span 10.0 pips, and 158.17 sits 52.0 to 53.0 pips beneath it. Neither new reading carries an as-of date. That is the eleventh consecutive slot in which we have asked for a dated 200-day and been given an undated one, and we are going to keep printing the absence rather than quietly using a number that is a function of a window nobody will name.
Three probabilities, one named feed, and a chain that is really one wire
For ten consecutive slots this desk has published that it could not find a December Fed probability with its feed named, and has refused to quote an unattributed percentage. Today we have one. Two separate pages, published 05:38 and 09:36, both state 84.4 per cent for a quarter point in December and both name the CME FedWatch tool explicitly; one adds 15.6 per cent for no change and zero for a cut, and the other adds that the same tool assigns no probability to a half point.
We are labelling it honestly rather than declaring the open item closed. Both pages credit the same crypto-market publisher as their origin, so two readings of one wire are not two independent readers. Alongside them sit 84.2 per cent with 15.8 per cent for no change from the 10:52 market outlook with no feed named, and “about 85 per cent” from the Reuters report, attributed to fed funds futures as of Monday’s close rather than to a tool. Three readings spanning 0.8 percentage points, with exactly one attribution chain that reaches a named feed, and that chain is one hop wide.
What is worth more than the number is the zero. The same tool that prices 84.4 per cent on a quarter point prices nothing on a half point, and that is a statement about the shape of the distribution, not a rounding. A reader who takes only the 84.4 away from today has taken the part Waller did not change.
The auction at 18:00 we still cannot pre-register
The Treasury sells 42 billion dollars of ten-year notes at 18:00 UTC, settling Monday 16 November, with 25 billion dollars of thirty-years on Thursday. Yesterday’s work order asked this desk to obtain a when-issued level so the auction could carry a proper pre-registration. We tried four routes this morning and have none. We are not going to write a test against a level we do not have, and we are not going to write one against a rolling average either — that is precisely the error we published and scored twelve hours ago, when a threshold set at a 65.9 per cent average was measured on auction day against a yardstick that had moved to 62.7.
What we can give you is the cash level and its spread. The ten-year slipped to around 4.855 per cent early in London on one reading and shows 4.8640 on a live ticker with no timestamp — 0.9 basis points apart, which is the kind of gap two differently-timed readings produce and not a dispute. Monday’s close was around 4.88 on the two-year at 4.56. So the auction prices into a ten-year that has firmed about two and a half basis points since Monday, into a CPI print four and a half hours earlier, and into a Fed governor who has publicly attached a 50 basis point branch to that print. If you run a blackout around the auction, note that the information event is not the auction.
What this does not tell you
It does not tell you the core number. Nothing here is a forecast of the print, and the only forecast we have relayed is a consensus of 0.2 per cent from pages that do not name their contributing houses.
It does not tell you that the half point is likely. One governor said it should be considered under a stated condition; that is not a committee, and the committee that held in October did so by ten votes to two. Waller’s own words are a conditional about what he would support, and the 0.3 per cent figure is Reuters’ wording for the condition rather than text inside his quotation marks — a distinction we are keeping because it is the difference between a threshold he set and a threshold a reporter inferred, and we cannot settle which from the pages we read.
It does not give you a time for the remarks. Three readers place them in London this morning and none states a clock time, so we have published no observation time for the speech itself and have derived every interval here from the actual clock.
It does not give you a when-issued level for the ten-year, a dated 200-day moving average, or a December probability from two genuinely independent feeds. Those are three things we looked for this morning and did not find, and we would rather you knew which three.
